feat: add mamba and dynamic chunking related code and test code

scripts/classification/DC_PatchTST.sh

@@ -0,0 +1,142 @@
+export CUDA_VISIBLE_DEVICES=0
+
+model_name=DC_PatchTST
+
+
+
+# DC_PatchTST specific parameters
+d_model_stage0=64        # Stage 0 dimension (D0)
+depth_enc0=1             # Stage 0 Mamba2 encoder depth  
+depth_enc1=1             # Stage 1 Mamba2 encoder depth
+target_ratio0=0.25       # Target compression ratio for stage 0
+target_ratio1=0.25        # Target compression ratio for stage 1
+
+# EthanolConcentration dataset
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/EthanolConcentration/ \
+  --model_id EthanolConcentration \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 8 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 8 \
+  --dropout 0.1 \
+  --activation gelu \
+  --des 'DC_PatchTST_Exp' \
+  --itr 1 \
+  --learning_rate 0.0002 \
+  --train_epochs 100 \
+  --patience 10 \
+  --d_model_stage0 $d_model_stage0 \
+  --depth_enc0 $depth_enc0 \
+  --depth_enc1 $depth_enc1 \
+  --target_ratio0 $target_ratio0 \
+  --target_ratio1 $target_ratio1
+
+# FaceDetection dataset  
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/FaceDetection/ \
+  --model_id FaceDetection \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 8 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 8 \
+  --dropout 0.1 \
+  --activation gelu \
+  --des 'DC_PatchTST_Exp' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --d_model_stage0 $d_model_stage0 \
+  --depth_enc0 $depth_enc0 \
+  --depth_enc1 $depth_enc1 \
+  --target_ratio0 $target_ratio0 \
+  --target_ratio1 $target_ratio1
+
+# Handwriting dataset
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/Handwriting/ \
+  --model_id Handwriting \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 8 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 8 \
+  --dropout 0.1 \
+  --activation gelu \
+  --des 'DC_PatchTST_Exp' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --d_model_stage0 $d_model_stage0 \
+  --depth_enc0 $depth_enc0 \
+  --depth_enc1 $depth_enc1 \
+  --target_ratio0 $target_ratio0 \
+  --target_ratio1 $target_ratio1
+
+# Heartbeat dataset
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/Heartbeat/ \
+  --model_id Heartbeat \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 8 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 8 \
+  --dropout 0.1 \
+  --activation gelu \
+  --des 'DC_PatchTST_Exp' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --d_model_stage0 $d_model_stage0 \
+  --depth_enc0 $depth_enc0 \
+  --depth_enc1 $depth_enc1 \
+  --target_ratio0 $target_ratio0 \
+  --target_ratio1 $target_ratio1
+
+# JapaneseVowels dataset
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/JapaneseVowels/ \
+  --model_id JapaneseVowels \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 8 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 8 \
+  --dropout 0.1 \
+  --activation gelu \
+  --des 'DC_PatchTST_Exp' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --d_model_stage0 $d_model_stage0 \
+  --depth_enc0 $depth_enc0 \
+  --depth_enc1 $depth_enc1 \
+  --target_ratio0 $target_ratio0 \
+  --target_ratio1 $target_ratio1

scripts/classification/vanillaMamba_classification.sh

@@ -0,0 +1,259 @@
+#!/bin/bash
+
+# vanillaMamba Classification Training Script for Multiple Datasets
+export CUDA_VISIBLE_DEVICES=0
+
+model_name=vanillaMamba
+
+# Create results directory if it doesn't exist
+mkdir -p ./results
+
+# UWaveGestureLibrary dataset (seq_len=315, enc_in=3) - use Copy1 config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/UWaveGestureLibrary/ \
+  --model_id UWaveGestureLibrary \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 315 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 128 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_UWaveGestureLibrary' \
+  --itr 1 \
+  --learning_rate 0.002 \
+  --train_epochs 150 \
+  --patience 30 \
+  --revin 0 | tee ./results/vanillaMamba_UWaveGestureLibrary.log
+
+# EthanolConcentration dataset (seq_len=1751, enc_in=3) - use Copy1 config
+python -u run.py \
+  --task_name classification \
+  --is_training 3 \
+  --root_path ./dataset/EthanolConcentration/ \
+  --model_id EthanolConcentration \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 1751 \
+  --enc_in 4 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_EthanolConcentration' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 200 \
+  --patience 30 \
+  --revin 0 | tee ./results/vanillaMamba_EthanolConcentration.log
+
+# Handwriting dataset (seq_len=152, enc_in=3) - use Copy1 config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/Handwriting/ \
+  --model_id Handwriting \
+  --model $model_name \
+  --data UEA \
+  --e_layers 4 \
+  --batch_size 64 \
+  --seq_len 152 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_Handwriting' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 200 \
+  --patience 30 \
+  --revin 0 | tee ./results/vanillaMamba_Handwriting.log
+
+# JapaneseVowels dataset (seq_len=29, enc_in=12) - use Copy1 config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/JapaneseVowels/ \
+  --model_id JapaneseVowels \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 64 \
+  --seq_len 29 \
+  --enc_in 12 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_JapaneseVowels' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 | tee ./results/vanillaMamba_JapaneseVowels.log
+
+# PEMS-SF dataset (seq_len=144, enc_in=963) - use Copy1 config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/PEMS-SF/ \
+  --model_id PEMS-SF \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 16 \
+  --seq_len 144 \
+  --enc_in 963 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_PEMS-SF' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 150 \
+  --patience 30 \
+  --revin 0 | tee ./results/vanillaMamba_PEMS-SF.log
+
+# Heartbeat dataset (seq_len=405, enc_in=61) - use original config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/Heartbeat/ \
+  --model_id Heartbeat \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 64 \
+  --seq_len 405 \
+  --enc_in 61 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_Heartbeat' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 150 \
+  --patience 10 \
+  --revin 0 | tee ./results/vanillaMamba_Heartbeat.log
+
+# FaceDetection dataset (seq_len=62, enc_in=144) - use original config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/FaceDetection/ \
+  --model_id FaceDetection \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 64 \
+  --seq_len 62 \
+  --enc_in 144 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_FaceDetection' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --revin 0 | tee ./results/vanillaMamba_FaceDetection.log
+
+# SelfRegulationSCP1 dataset (seq_len=896, enc_in=6) - use original config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/SelfRegulationSCP1/ \
+  --model_id SelfRegulationSCP1 \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 64 \
+  --seq_len 896 \
+  --enc_in 6 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_SelfRegulationSCP1' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --revin 0 | tee ./results/vanillaMamba_SelfRegulationSCP1.log
+
+# SelfRegulationSCP2 dataset (seq_len=1152, enc_in=7) - use original config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/SelfRegulationSCP2/ \
+  --model_id SelfRegulationSCP2 \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 64 \
+  --seq_len 1152 \
+  --enc_in 7 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_SelfRegulationSCP2' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --revin 0 | tee ./results/vanillaMamba_SelfRegulationSCP2.log
+
+# SpokenArabicDigits dataset (seq_len=93, enc_in=13) - use original config
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/SpokenArabicDigits/ \
+  --model_id SpokenArabicDigits \
+  --model $model_name \
+  --data UEA \
+  --e_layers 3 \
+  --batch_size 64 \
+  --seq_len 93 \
+  --enc_in 13 \
+  --d_model 128 \
+  --d_state 64 \
+  --d_conv 4 \
+  --expand 2 \
+  --headdim 64 \
+  --dropout 0.1 \
+  --des 'vanillaMamba_SpokenArabicDigits' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 10 \
+  --revin 0 | tee ./results/vanillaMamba_SpokenArabicDigits.log

scripts/classification/xPatch_SparseChannel-Copy1.sh

@@ -0,0 +1,145 @@
+#!/bin/bash
+
+# xPatch_SparseChannel Classification Training Script for Multiple Datasets
+export CUDA_VISIBLE_DEVICES=0
+
+model_name=xPatch_SparseChannel
+
+# Create results directory if it doesn't exist
+mkdir -p ./results
+
+
+
+# UWaveGestureLibrary dataset (seq_len=315, enc_in=3, k_graph=3)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/UWaveGestureLibrary/ \
+  --model_id UWaveGestureLibrary \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 315 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_UWaveGestureLibrary' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 3 | tee ./results/xPatch_SparseChannel_UWaveGestureLibrary.log
+
+
+
+# EthanolConcentration dataset (seq_len=1751, enc_in=3, k_graph=3)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/EthanolConcentration/ \
+  --model_id EthanolConcentration \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 1751 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_EthanolConcentration' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 3 | tee ./results/xPatch_SparseChannel_EthanolConcentration.log
+
+# Handwriting dataset (seq_len=152, enc_in=3, k_graph=3)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/Handwriting/ \
+  --model_id Handwriting \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 152 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_Handwriting' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 3 | tee ./results/xPatch_SparseChannel_Handwriting.log
+
+# JapaneseVowels dataset (seq_len=29, enc_in=12, k_graph=8)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/JapaneseVowels/ \
+  --model_id JapaneseVowels \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 29 \
+  --enc_in 12 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_JapaneseVowels' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 8 | tee ./results/xPatch_SparseChannel_JapaneseVowels.log
+
+
+
+# PEMS-SF dataset (seq_len=144, enc_in=963, k_graph=8)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/PEMS-SF/ \
+  --model_id PEMS-SF \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 16 \
+  --seq_len 144 \
+  --enc_in 963 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_PEMS-SF' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 8 | tee ./results/xPatch_SparseChannel_PEMS-SF.log

scripts/classification/xPatch_SparseChannel.sh

@@ -1,10 +1,66 @@
 #!/bin/bash
 
-# xPatch_SparseChannel Classification Training Script for FaceDetection Dataset
+# xPatch_SparseChannel Classification Training Script for Multiple Datasets
 export CUDA_VISIBLE_DEVICES=0
 
 model_name=xPatch_SparseChannel
 
+# Create results directory if it doesn't exist
+mkdir -p ./results
+
+# Heartbeat dataset (seq_len=405, enc_in=61, k_graph=8)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/Heartbeat/ \
+  --model_id Heartbeat \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 405 \
+  --enc_in 61 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_Heartbeat' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 5 \
+  --revin 0 \
+  --k_graph 8 | tee ./results/xPatch_SparseChannel_Heartbeat.log
+
+# UWaveGestureLibrary dataset (seq_len=315, enc_in=3, k_graph=3)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/UWaveGestureLibrary/ \
+  --model_id UWaveGestureLibrary \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 315 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_UWaveGestureLibrary' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 3 | tee ./results/xPatch_SparseChannel_UWaveGestureLibrary.log
+
+# FaceDetection dataset (seq_len=62, enc_in=144, k_graph=8)
 python -u run.py \
   --task_name classification \
   --is_training 1 \
@@ -12,21 +68,202 @@ python -u run.py \
   --model_id FaceDetection \
   --model $model_name \
   --data UEA \
-  --e_layers 3 \
+  --e_layers 2 \
   --batch_size 64 \
   --seq_len 62 \
   --enc_in 144 \
   --d_model 128 \
   --d_ff 256 \
-  --n_heads 8 \
+  --n_heads 16 \
   --patch_len 16 \
   --stride 8 \
-  --moving_avg 25 \
   --dropout 0.1 \
   --des 'xPatch_SparseChannel_FaceDetection' \
   --itr 1 \
   --learning_rate 0.0005 \
   --train_epochs 100 \
   --patience 5 \
-  --revin 1 \
-  --k_graph 8
+  --revin 0 \
+  --k_graph 8 | tee ./results/xPatch_SparseChannel_FaceDetection.log
+
+# EthanolConcentration dataset (seq_len=1751, enc_in=3, k_graph=3)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/EthanolConcentration/ \
+  --model_id EthanolConcentration \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 1751 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_EthanolConcentration' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 3 | tee ./results/xPatch_SparseChannel_EthanolConcentration.log
+
+# Handwriting dataset (seq_len=152, enc_in=3, k_graph=3)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/Handwriting/ \
+  --model_id Handwriting \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 152 \
+  --enc_in 3 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_Handwriting' \
+  --itr 1 \
+  --learning_rate 0.001 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 3 | tee ./results/xPatch_SparseChannel_Handwriting.log
+
+# JapaneseVowels dataset (seq_len=29, enc_in=12, k_graph=8)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/JapaneseVowels/ \
+  --model_id JapaneseVowels \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 29 \
+  --enc_in 12 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_JapaneseVowels' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 8 | tee ./results/xPatch_SparseChannel_JapaneseVowels.log
+
+# SelfRegulationSCP1 dataset (seq_len=896, enc_in=6, k_graph=6)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/SelfRegulationSCP1/ \
+  --model_id SelfRegulationSCP1 \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 896 \
+  --enc_in 6 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_SelfRegulationSCP1' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 5 \
+  --revin 0 \
+  --k_graph 6 | tee ./results/xPatch_SparseChannel_SelfRegulationSCP1.log
+
+# SelfRegulationSCP2 dataset (seq_len=1152, enc_in=7, k_graph=7)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/SelfRegulationSCP2/ \
+  --model_id SelfRegulationSCP2 \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 1152 \
+  --enc_in 7 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_SelfRegulationSCP2' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 5 \
+  --revin 0 \
+  --k_graph 7 | tee ./results/xPatch_SparseChannel_SelfRegulationSCP2.log
+
+# SpokenArabicDigits dataset (seq_len=93, enc_in=13, k_graph=8)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/SpokenArabicDigits/ \
+  --model_id SpokenArabicDigits \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 64 \
+  --seq_len 93 \
+  --enc_in 13 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_SpokenArabicDigits' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 5 \
+  --revin 0 \
+  --k_graph 8 | tee ./results/xPatch_SparseChannel_SpokenArabicDigits.log
+
+# PEMS-SF dataset (seq_len=144, enc_in=963, k_graph=8)
+python -u run.py \
+  --task_name classification \
+  --is_training 1 \
+  --root_path ./dataset/PEMS-SF/ \
+  --model_id PEMS-SF \
+  --model $model_name \
+  --data UEA \
+  --e_layers 2 \
+  --batch_size 16 \
+  --seq_len 144 \
+  --enc_in 963 \
+  --d_model 128 \
+  --d_ff 256 \
+  --n_heads 16 \
+  --patch_len 16 \
+  --stride 8 \
+  --dropout 0.1 \
+  --des 'xPatch_SparseChannel_PEMS-SF' \
+  --itr 1 \
+  --learning_rate 0.0005 \
+  --train_epochs 100 \
+  --patience 30 \
+  --revin 0 \
+  --k_graph 8 | tee ./results/xPatch_SparseChannel_PEMS-SF.log